conditioning

conditioning My heart races and my palms sweat during my first attempt to drive again after a traumatic road accident. Alternatively, having discovered that a new joke goes down well at work, I find myself retelling it ad nauseam. These two scenarios are examples of different forms of conditioning. The first is an example of classical conditioning, and involves learning about the predictive relationship between stimuli. As a result of the car accident, the interior of a car, let alone the touch of steering wheel, has become a signal for a traumatic event and thereby elicits a fear reaction through the activation of my autonomic nervous system. By contrast, my new found humour is an example of ‘instrumental’ or ‘operant’ conditioning. In this case, I have learned about the causal relationship between an action, the telling of the joke, and the apparent attention and interest that it elicits in my friends and colleagues, which only serves to reward this tedious behaviour.

Classical conditioning is often referred to as ‘Pavlovian’ because this form of learning was discovered by the renowned Russian physiologist, Pavlov, in his experiments on the neural control of digestion at the end of the nineteenth century. As is well known, Pavlov signalled the presentation of food to his hungry dogs by turning on a stimulus, such a bell, some seconds before the delivery of each meal. Although the bell initially produced little more than orientation towards its source, after a number of pairings with the food this stimulus began to elicit novel behaviour. As soon as the signal came on, the dogs approached the location of the food and started salivating copiously. The occurrence of the responses depended, or were conditional, upon experience of the predictive relationship between the signal and the food, and thus came to be known as ‘conditioned’. Correspondingly, the signal is called a conditioned stimulus, because its property also depends upon learning about the predictive relationship. By contrast, the food is an unconditioned stimulus, because the salivation that it elicits, the unconditioned response, does not depend upon the learning experience. Pavlov also referred to the food as a reinforcer, as it is the event responsible for strengthening the conditioned response. Although it was originally thought that simple pairings of a conditioned stimulus and a reinforcer are sufficient for conditioning, we now know that only signals that are informative about the occurrence of the reinforcer become conditioned. Moreover, conditioning is not always a simple, automatic and non-conscious process and, in certain cases, only occurs in humans when they are already aware of the relationship between signal and reinforcer.

The salivation elicited by the signals for food is an appetitive conditioned response because the reinforcer, the food, is attractive. By contrast, my hypothetical fear response to the car is an example of aversive or defensive conditioning, because the reinforcer in this case, the accident, is noxious and distressing. Pavlovian conditioning affects a gamut of response and behaviour systems, from the sexual evaluation of members of the opposite sex to food preference and aversions. Moreover, this form of conditioning also plays a role not only in behavioural responses but also in the regulatory systems of the body. For example, if drinking a fluid with a particular flavour signals an infusion of glucose into the stomach of hungry rats, that flavour will, in future, reduce blood sugar level in anticipation of the glucose load.

The experimental study of ‘instrumental’ conditioning also started over 100 years ago, but in this case by an American comparative psychologist, Thorndike, who was interested in comparing the learning capacities of different species of animal. Thorndike studied the rate at which a variety of animals learned to operate a latch in order to escape from a cage to eat some food placed outside. These instrumental tasks were subsequently refined over the succeeding decades, most notably by the behaviourist psychologist, Skinner. As in the case of Pavlovian conditioning, the food acted as a reinforcer to strengthen the conditioned response, the operation of the latch, but in the instrumental case through a positive causal relationship between the response and reinforcer. In contrast to Pavlovian conditioning, however, aversive or noxious stimuli cannot act as instrumental reinforcers through a positive relationship with a response. Indeed, when a response causes an aversive outcome, the behaviour is suppressed or punished. For an aversive event, such as a road accident, to reinforce the appropriate instrumental response (careful and defensive driving) the response has to prevent the event happening and thereby allow us to escape or avoid dangerous and unpleasant situations.

There are two sorts of learning process underlying instrumental conditioning. The first process establishes response habits through the acquisition of a connection between an eliciting stimulus and the response. For example, enhancement of the limb muscle reflexes involved in the movements that the rat must make to reach for the latch can be conditioned by arranging for an appropriate change to be reinforced by the delivery of food to a hungry animal. This simple stimulus–response development clearly plays a role in the acquisition of motor skills. Other learning processes are involved in more complex forms of instrumental conditioning, which support goal-directed actions based upon knowledge of the causal relationship between the action and the outcome that it achieves. This type of instrumental conditioning operates when one explicitly plans a course of action to achieve a specific goal.

In summary, the two forms of conditioning, Pavlovian and instrumental, reflect the processes by which we and other animals learn to adjust our behaviour to the predictive and causal structure of our environment. The fact that, in one form or another, both types of conditioning are to be found throughout the animal kingdom, from relatively simple invertebrates to ourselves, is a testimony to their ubiquitous and important adaptive function.

A. Dickinson

Bibliography

Dworkin, B. R. (1993). Learning and physiological regulation. The University of Chicago Press, Chicago.
Mackintosh, N. J. (1983). Conditioning and associative learning. Oxford University Press, New York.

See also Pavlov, Ivan.